De Novo Assembly Algorithms

Abstract

In an ideal case, an assembly algorithm should merge overlapped reads to one long continuous sequence, called contig, which is a chromosome in the primitive genome. But due to sequencing errors and the existence of unsequenced parts, contigs gained from the assembly algorithm are not complete enough to form chromosomes. Even with high coverage, there is still a non-zero probability for the existence of unsequenced parts and sequencing errors. The ability of the assembler to form contigs is also affected by repeated regions in the genome. As shown in Fig. 3.3 in the previous chapter, two parts of different repeat areas are mapped to one in the assembler because of the weakness of repeat detection in the assembler. Figure 4.1 shows how a typical assembly algorithm works in overlap detection and contig generation phases.